SMLL: Using 200MB of Neural Network to Save 400 Bytes

(frankchiarulli.com)

14 points | by fcjr 19 hours ago

1 comments

  • msephton 8 hours ago

    No mention of decompression speed and validation, or did I miss something?

    • savalione 8 hours ago

      It's in the post: Benchmarks -> Speed

      tl;dr: SMLL is approximately 10,000x slower than Gzip